/status

name: status
description: Check whether ai-memory is capturing this session and its memory tools can reach the server. Run it when memory looks missing or after installing or upgrading ai-memory.
disable-model-invocation: true

ai-memory status

Report on the ai-memory install this session uses. Only report: do not change any file, setting or hook, and do not run any command other than the two below.

1. Run `ai-memory --version`. If the command is not found, say that the ai-memory binary is not installed (or not on `PATH`), that this session is therefore not captured while the memory tools may still answer from an ai-memory server, and point to <https://github.com/akitaonrails/ai-memory#install>. Stop there. 2. Run `ai-memory doctor` in the current project directory.

Summarize the doctor output in a few short lines:

• whether the server answered, and which server URL it used;
• the workspace and project this directory maps to;
• capture coverage for this agent, and any agent it flags as uncaptured;
• the hook spool backlog, if any events are waiting;
• duplicate wiring: the ai-memory plugin and settings-file hooks (or

installed skills) both active, and which one doctor suggests keeping;

• for Codex, any ai-memory hook Codex skips as `untrusted` or `modified`

(the user trusts them in `/hooks`);

• the ai-memory plugin version against the binary version, when they differ.

Quote problems as doctor states them, and leave every fix to the user.